Skip to main content
NCBI home page
Proceedings of the National Academy of Sciences of the United States of America logoLink to Proceedings of the National Academy of Sciences of the United States of America
. 1998 Oct 27;95(22):12934–12939. doi: 10.1073/pnas.95.22.12934

Imaging of H217O distribution in the brain of a live rat by using proton-detected 17O MRI

Itamar Ronen *, Hellmut Merkle †, Kamil Ugurbil †, Gil Navon *,‡
PMCID: PMC23660  PMID: 9789018

Abstract

Imaging of H217O has a number of important applications. Mapping the distribution of H217O produced by oxidative metabolism of 17O-enriched oxygen gas may lead to a new method of metabolic functional imaging; regional cerebral blood flow also can be measured by measuring the H217O distribution after the injection of 17O-enriched physiological saline solution. Previous studies have proposed a method for indirect detection of 17O. The method is based on the shortening of the proton T2 in H217O solutions, caused by the residual 17O-1H scalar coupling and transferred to the bulk water via fast chemical exchange. It has been shown that the proton T2 of H217O solutions can be restored to that of H216O by irradiating the resonance frequency of the 17O nucleus. The indirect 17O image thus is obtained by taking the difference between two T2-weighted spin-echo images: one acquired after irradiation of the 17O resonance and one acquired without irradiation. It also has been established that, at relatively low concentrations of H217O, the indirect method yields an image that quantitatively reflects the H217O distribution in the sample. The method is referred to as PRIMO (proton imaging of oxygen). In this work, we show in vivo proton images of the H217O distribution in a rat brain after an i.v. injection of H217O-enriched physiological saline solution. Implementing the indirect detection method in an echo-planar imaging sequence enabled obtaining H217O images with good spatial and temporal resolution of few seconds.

MRI is an emerging technique for functional imaging of the brain (f-MRI). The most commonly used method is blood oxygenation level dependent f-MRI, which exploits the difference in magnetic properties between the diamagnetic oxyhemoglobin [HbO2] and the paramagnetic deoxyhemoglobin [Hb] (1–4). Local changes in the content of Hb give rise to local variations in signal intensity of a gradient-echo image. However, interpretation of the changes in signal intensity is highly complex because of the multitude of factors that influence these changes: the overall net change in signal intensity is a combination of the decrease in Hb that results from changes in local blood flow and local blood volume, caused by increased influx of freshly oxygenated blood, whereas oxygen consumption presumably is elevated commensurately less. The complexity of the analysis of the blood oxygenation level dependent effect is the main incentive for devising new f-MRI methods that would supply more direct evidence for local oxidative metabolism in the brain.

Another functional imaging method used for following oxidative metabolism in the brain is 15O-positron-emission tomography (5, 6). The subject inhales molecular oxygen enriched with 15O2, which is converted into H215O in regions in which oxidative metabolism takes place. However, 15O-positron-emission tomography is unable to distinguish between 15O2 in oxyhemoglobin and 15O incorporated in H215O molecules. The advantage of 17O MRI over 15O-positron-emission tomography would be that 17O cannot be detected as molecular 17O2 dissolved in the blood, which is paramagnetic, or as 17O2 bound to hemoglobin because of the extremely broad 17O resonance caused by quadrupolar relaxation enhanced by the slow motion of the macromolecule. 17O becomes detectable by NMR only after it is converted to H217O, and, therefore, 17O MRI images of active regions of the brain would reflect oxygen consumption directly and would not reflect additional mechanisms, such as changes in blood volume and blood flow.

The potential of H217O as a tracer of oxygen consumption already has been pointed out in several works (8–19)§. Direct 17O-NMR has been efficient in providing data of whole brain oxygen consumption rate of a cat (14) after inhalation of molecular oxygen enriched with 17O2, and a rough distribution of oxygen consumption rate in the cat brain was achieved via direct 17O-MRI. However, 17O is highly unsuitable for MRI because of its low sensitivity and short relaxation times, making 17O f-MRI impractical. An alternative way of using H217O was suggested by Hopkins (16–19) and Arai (9), exploiting the characteristics of H217O as a contrast agent. The transverse relaxation time of water protons in H217O solutions is known to be shorter than that of water protons of H216O. Meiboom (20) showed that the reason for this shortening is the residual line broadening caused by the 1H-17O scalar coupling. The scalar splitting (J = 92.5 Hz; ref. 21) is coalesced almost completely as a result of fast chemical exchange with H216O and the quadrupolar relaxation of the 17O nucleus. Nonetheless, the shortening in the water proton T2 is significant even when H217O concentration is as low as the natural abundance (0.037%). Kwong et al. (18) have shown an in vivo application of the use of H217O as a contrast agent, in which a dog was injected with a bolus of H217O-enriched physiological saline solution, and a series of echo-planar images (EPI) showed the contrast created by the diffusion of the bolus in the dog’s brain. The main disadvantage of this method is related to the need for a “baseline” image that would reflect the anatomic image without contrast. Such an image should be acquired in a separate imaging session before the injection of the contrast agent or after complete washout. The use of different imaging sessions inevitably introduces time-dependent errors, thus making the task of accurate quantitation of local oxygen consumption almost impossible.

Previous studies (22, 23) have presented a method for proton-detected 17O-MRI that combines the advantages of 1H-MRI with sensitive detection of 17O. The method is based on the finding that RF irradiation at the 17O resonance frequency removes the residual 1H-17O scalar coupling, thus increasing the water proton T2. Hence, when a T2-weighted spin-echo image is subtracted from a similar image in which the 17O resonance has been irradiated, the resulting difference image is a proton image that reflects the distribution of H217O in the observed sample. This method can be accomplished by using either two separate coils tuned to the 1H and 17O frequencies (22, 23) or a double tuned coil (24, 25). It also has been shown that, when time-to-echo (TE) values are of the order of the water proton T2, the resulting image achieves its maximum intensity, and, for low H217O, the image quantitatively reflects the H217O distribution in the observed object. In this work, we present an in vivo demonstration of proton-detected 17O-MRI. We show that proton-detected 17O-MRI is capable of providing the in vivo distribution of low concentrations of H217O in a rat brain, introduced via i.v. injection of 0.9% saline solution that contained H217O. We show that PRIMO (proton imaging of oxygen-17) is readily applicable both in multiple shot imaging techniques such as spin-echo PRESS and in fast one-shot techniques such as spin-echo EPI.

THEORETICAL BACKGROUND

The theoretical basis of the PRIMO experiment was given in previous publications (22, 23, 25). A brief review of the method, together with some extensions, is given here. As was shown by Meiboom (20), water proton T2 is affected by the presence of H217O molecules because of the 1H-17O spin–spin coupling and proton chemical exchange between H217O and H216O molecules. An accurate measurement of the 17O-1H coupling constant yielded a value of 92.5 Hz (21). Denoting T2(16) and T2(17) as the proton transverse relaxation times of H216O and H217O, respectively, and assuming fast proton exchange between these two species, the proton transverse relaxation rate of water containing H217O at a molar fraction P is given by

graphic file with name M1.gif 1

where the term P(1/T2(17) − 1/T2(16)) is the contribution to the water proton relaxation resulting from the presence of H217O. Thus, the spin-echo signal of water containing H217O at a molar fraction P as a function of TE is given by

graphic file with name M2.gif 2

Assuming that the 17O irradiation completely abolishes the effect caused by the 1H-17O spin-spin coupling and that the contribution of the 1H-17O dipolar interaction to the water proton relaxation is negligible, the proton spin echo signal of a H217O solution on 17O irradiation is given by

graphic file with name M3.gif 3

The difference between the two signals is

graphic file with name M4.gif 4
graphic file with name M5.gif

For the case in which the value of the exponent in the square brackets is much smaller than unity (see below),

graphic file with name M6.gif 5

The normalized difference signal is given by

graphic file with name M7.gif 6

Because this term is much smaller than unity in all practical applications, the derivation of Eq. 5 is justified.

The contribution of H217O molecules to the water proton transverse relaxation rate in the presence of H217O has been shown23 to obey the relation

graphic file with name M8.gif 7

where τ is the characteristic proton exchange lifetime in H217O, J is the 17O-1H scalar coupling constant, and T1 is the longitudinal relaxation time of the 17O nucleus. Combining Eqs. 6 and 7 gives

graphic file with name M9.gif 8

Meiboom (20) showed that, in pure water (neutral pH), τ is 1.8 × 10−3 s and that T1 is 4.4 × 10−3 s. In biological tissues, the 17O longitudinal relaxation time is not considerably shortened whereas τ, which is sensitive to the presence of buffers, might be shortened (7, 26, 27). In that case, Eq. 8 simplifies to

graphic file with name M10.gif 9

One should note that the normalized difference signal is independent of the water proton T2 but depends on the water proton exchange rate.

MATERIALS AND METHODS

The experiments were performed by using a 4.7T Varian MRI spectrometer. The coil used for the experiment was a combination of a single 1H surface coil (20 mm diameter) tuned to 200 MHz and a quadrature coil (two 22 mm diameter coils with partial overlap) tuned to 27.13 MHz for the 17O, intended to minimize the power deposition that might have been caused by irradiation at the 17O resonance frequency. The power used for the 17O irradiation was ≈3 W (90° pulse = 120 μs). The pulse sequence used for the first imaging experiment was decoupled point-resolved spectroscopy (DEPRESS) (decoupled PRESS), a modification of the spin-echo-based adiabatic PRESS imaging in which a nonselective adiabatic excitation pulse was followed by two slice selective adiabatic refocusing pulses separated by delays. Each k-space row was scanned twice, one after the other, with and without irradiation at the 17O resonance frequency. These rows later were combined and processed separately to give the irradiated image and the image without irradiation, respectively. The PRIMO image was obtained by subtracting the two images and dividing the difference image by the irradiated image. The pulse sequence used in the second imaging experiment was DEPRESS-EPI. It consisted of two contiguous T2-weighted EPI sequences. During TE of the first EPI shot, the 17O was irradiated with a CW field at the 17O resonance frequency whereas there was no irradiation during the second EPI shot. The excitation pulse of the first EPI shot was a 45° BIR4 adiabatic pulse, and the excitation pulse of the second EPI shot was a 90° adiabatic pulse, so that the initial transverse magnetizations used for both EPI shots were approximately equal in magnitude. The 90° adiabatic pulse was applied immediately at the end of the acquisition period (60 ms) that followed the 45° pulse. A relaxation delay was inserted after the image acquisition that followed the 90° pulse. Fine tuning of the method can be done by running the 45–90 EPI sequence without irradiation at the 17O frequency and by using any deviation from zero difference as a correction factor. The advantage of this method is that the two images that are subtracted from each other are obtained almost simultaneously; thus, it minimizes time-dependent errors such as motion artifacts.

Water enriched with H217O was purchased from Enritech (Tel Aviv) and Isotec (Miamisburg, OH). The phantom was an array of seven 5-mm tubes arranged in a concentric pattern. Each tube contained 7% BSA solution with a different H217O concentration, varying from natural abundance (0.037%) to 0.56%. The tubes were plugged in a cylindrical plastic holder that was inserted in a large plastic tube that contained a 1:1 mixture of H2O and D2O. Proton T2 values of the BSA solutions in the tubes were ≈0.4 s.

The animal model in both experiments was a 300-g Sprague–Dawley rat. The rat used for the DEPRESS experiment was anaesthetized initially with 0.3 cc of a mixture of 10 mg/ml acepromazine, 20 mg/ml xylacine, and 100 mg/ml ketamine, which was administered i.m. Throughout the rest of the experiment, the rat was anaesthetized with 25-μl doses of the same mixture diluted 4-fold, administered i.v. every 5 min. Blood pressure and heart rate were monitored continuously to ensure that the rat was alive throughout the experiment. Proton T2 in the brain ranged between 55 ms (gray and white matter) and 150 ms [cerebrospinal fluid (CSF)]. The rat was injected i.v. with three successive doses of 1 cc of 0.9% physiological saline solution that contained 46.5% H217O. During injection, the H217O concentration was monitored via direct 17O-NMR. When steady state was reached, the H217O concentration was calculated by comparing the intensity of the 17O peak after the injection to that of the 17O peak at natural abundance, measured at the beginning of the experiment. At each 17O concentration, four imaging sessions were performed with four different TE values. The whole series of experiment took ≈7 hours.

The rat used for the DEPRESS-EPI experiment was tracheotomized and anaesthetized via inhalation of 2–3.5% isofluorane combined with a 7:3 mixture of oxygen and nitrous oxide. Occasional blood samples were drawn from a femoral arterial line for blood gas analysis, performed by using an IL1304 blood gas analyzer (Instrumentation Laboratory, Lexington, MA.). Heart rate and blood oxygenation level were monitored constantly by using a pulse oxymeter (Nonin, Plymouth, MN) secured to the rat’s tail. Physiological saline solution (0.9%) enriched with 10% H217O was administered i.v. through a femoral venous line. An 17O NMR spectrum was acquired before the injection to record the intensity of the 17O peak at natural abundance. Then, a series of 100 pairs of DEPRESS-EPI images was launched. Each image was acquired in 60 ms, and a pair (with and without 17O irradiation) was acquired in 120 ms. The interval between consecutive pairs was 5 s. During the first 50 experiments, a dose of 3 ml of the 0.9% physiological saline solution was injected slowly, so that the rate of injection was roughly constant. During the last 50 pairs of images, no injection took place. Immediately after the experiments were over, a direct 17O NMR measurement was performed. The procedure was repeated three times consecutively, so that the total amount of 0.9% physiological saline solution injected to the rat was ≈10 ml. This relatively large volume of solution was used because of the low level of enrichment of the injected solution. The experiment took ≈1 hour.

RESULTS

Phantom Results.

In Fig. 1, DEPRESS images of the phantom at TE = 0.4 s are given. Fig. 1a is the image acquired without 17O irradiation, and Fig. 1b is the one acquired with 17O irradiation. Fig. 1c is the difference between Fig. 1b and Fig. 1a. The intensity of the image is modulated not only by the different concentrations of H217O but also by the sensitivity profile of the surface coil in signal reception (because of the inhomogeneity of B1 generated by the surface coil) across the phantom. The intensity modulation caused by B1 inhomogeneity is corrected in Fig. 1d, the PRIMO image, which is the difference image (Fig. 1c) divided by the irradiated image (Fig. 1b). Fig. 1e is a schematic drawing of the phantom in which the H217O concentrations in the various tubes are given. Fig. 2 is a plot of the average intensity of the normalized difference signal measured in each tube on the PRIMO image as a function of the H217O content of the tube. It should be noticed that even the tube that contained H217O in natural abundance (0.037%) gave rise to visible difference signal. The linear correlation between the normalized difference and the H217O concentration is maintained for a wide range of H217O concentrations that obey the restriction put on the exponent in Eq. 4 and breaks down when the H217O concentration is relatively high.

Figure 1.

Figure 1

Open in a new tab

DEPRESS images of the phantom. (a) Without 17O irradiation. (b) With 17O irradiation. (c) Difference image. (d) PRIMO image. (e) Schematic drawing of the phantom.

Figure 2.

Figure 2

Open in a new tab

Relative increase of the 1H spin echo signal measured in the various tubes of the phantom vs. the respective H217O concentration in the tubes.

In Vivo Results.

DEPRESS-PRIMO images. Fig. 3 a and b are typical coronal images of the rat’s brain obtained by using the DEPRESS sequence, obtained without and with 17O irradiation, respectively. Fig. 3c is the difference between the two images. CSF areas appear brighter than gray/white matter areas because the proton T2 of CSF water is considerably longer than that of gray/white matter water protons and the intensity of the difference signal is weighted by the T2 of H216O (Eq. 4). To remove the dependence of the intensity of the difference signal coming from different parts of the brain both on the local proton T2 value, as well as on the inhomogeneity of the B1 field generated by the surface coil, the image was normalized by dividing it by the image in Fig. 3b. This image, obtained after irradiation, reflects the water distribution in the brain weighted by the T2 of H216O (Eq. 3) where St=0 is weighted by B1 inhomogeneity. The intensity of the resulting PRIMO image is given by Eq. 8. The normalization procedure produces a distribution map of the H217O that is almost independent of the water proton T2 and depends almost exclusively on the water proton exchange rate, with some contribution of the 17O T1. Fig. 3d is the normalized image. It is evident that the contrast between the various brain areas is considerably lower than in Fig. 3c, as would be expected for an almost homogeneous molar fraction of H217O in the rat brain at steady state conditions. Still, the CSF areas appear brighter in the PRIMO image, probably because the water proton exchange rate in CSF, which is expected to be similar to the exchange rate in pure water, is slower than that in gray/white matter areas, where the water proton exchange rate may be enhanced by the presence of proteins. The fact that the water concentration in the gray/white matter area is lower than that in the CSF should not be reflected by the PRIMO image because in the normalization procedure the difference is divided by the water signal in the same voxel. Figs. 4 a–d are PRIMO images performed with TE = 0.1 s. This TE value is larger than the proton T2 of the water present in the cortex, which was ≈60 ms, and shorter than the CSF proton T2, which is ≈150 ms. The image in Fig. 4a was acquired before the injection of the 0.9% physiological saline solution and served as a baseline image. The image in Fig. 4b was taken after the rat was injected with the first i.v. injection of 1-cc dose of 0.9% physiological saline solution that contained 46% H217O. The image in Fig. 4c was taken after the second injection, and the image in Fig. 4d was taken after the third one. 17O-NMR measurements were performed before the first injection and after each injection of 0.9% physiological saline solution. Whole head H217O concentrations were calculated by comparing the 17O peak intensity after the injection to that of the peak at natural abundance, corresponding to 0.037% H217O. Fig. 5 shows the time course of the average PRIMO-DEPRESS intensities in two regions of the rat brain, one that includes mainly CSF and the other that includes mainly gray and white matter. To combine results taken at different TE values, we plot the relative enhancement divided by the TE. This value remained approximately constant for each tissue at a given H217O concentration, in spite of the large variation of the  TE values. Also shown in the figure is the 17O signal intensities from the whole brain taken during the injection of the 17O enriched 0.9% physiological saline solution.

Figure 3.

Figure 3

Open in a new tab

DEPRESS images of the rat brain. TE = 0.1 s., average H217O concentration over the whole brain as measured by 17O-NMR was ≈0.45%. (a) Without irradiation. (b) With irradiation. (c) Difference image. (d) PRIMO image.

Figure 4.

Figure 4

Open in a new tab

DEPRESS images of the rat brain at different H217O concentrations. (a) [H217O] = 0.037%. (b) [H217O] = 0.21%. (c) [H217O] = 0.36%. (d) [H217O] = 0.45%. Error estimate is calculated according to the SD of the PRIMO image as measured over a gray/white matter region and is equivalent to a concentration of ≈0.1% H217O.

Figure 5.

Figure 5

Open in a new tab

Time course of the H217O concentrations in the rat brain in the DEPRESS experiment. Solid line connects between direct 17O NMR measurements, and the various symbols refer to DEPRESS measurements. Empty symbols refer to CSF areas, and full symbols refer to gray/white matter areas. Squares, TE = 0.07 s; triangles, TE = 0.1 s; rhombi, TE = 0.15 s; circles, TE = 0.2 s. DEPRESS image at each TE value took 20 min, each direct 17O-NMR set of measurements during injections (longer periods) took 2 min, and each one of the shorter 17O-NMR measurements took ≈0.5 min. Owing to the death of the rat, only the measurements with TE = 0.07 s and TE = 0.1 s were taken.

Dynamic experiment: DEPRESS-EPI PRIMO images.

Real time measurements of H217O distribution in the rat brain were performed by using the DEPRESS-EPI imaging sequence. Each image was acquired in 60 ms, and the pair of images with and without 17O irradiation, including the two TE intervals (TE = 100 ms) took 320 ms. Pairs of images were acquired every 5 s. The experiment started with a direct measurement of the H217O concentration in the whole head by using 17O NMR. A 10% 17O-enriched 0.9% NaCl solution was injected i.v. into the femoral vein during 4.5 min (55 PRIMO images). Acquisition continued for an additional 4 min (45 PRIMO images), and then the 17O NMR measurement was repeated. This procedure was repeated three times. The amount of the injected 17O-enriched 0.9% physiological saline solution was 3 ml in the first and second injections and 4 ml in the third one. The results of the PRIMO signals as measured in two brain areas are given in Fig. 6. As can be seen in the figure, the PRIMO-detected H217O concentration continually increases during the injection of the enriched 0.9% physiological saline solution and slowly decreases after the termination of the injection because of equilibration of the H217O throughout the whole body. The results for the CSF and the gray/white matter are comparable in spite of the big difference in their proton T2. The remaining difference is probably caused by differences in the proton exchange rates and 17O T1 values in these tissues.

Figure 6.

Figure 6

Open in a new tab

Time course of the H217O concentrations in the rat brain as measured by the DEPRESS-EPI experiment performed with TE = 0.1 s. Solid line, DEPRESS-EPI signal from gray/white matter areas; dashed line, DEPRESS-EPI signal from CSF areas; triangles, direct 17O-NMR measurements.

DISCUSSION AND CONCLUSIONS

We have demonstrated that proton imaging of H217O (PRIMO) is a viable tool for mapping H217O distribution in a tissue and that the method can be applied for in vivo studies involving H217O. In our study, the relative increase in the spin-echo signal coming from cortex areas after 17O irradiation was ≈10% when the H217O concentration was ≈0.5%. This result was obtained by using TE values that maximized the absolute difference between the irradiated and nonirradiated signals. These facts suggest that proton-detected 17O-MRI is sensitive enough to detect small concentrations of H217O. Normalization of the difference image according to the image acquired after 17O irradiation eliminates both the variance in signal difference caused by B1 inhomogeneity and the dependence of the difference signal on the water proton T2. However the dependence of the PRIMO result from the water proton exchange time requires the calibration of the method for each tissue. A much larger relative increase could be achieved by using longer TE values, so that natural abundance of H217O became clearly visible. However, this came about in expense of lower signal-to-noise ratios.

We have shown previously that the overall sensitivity of proton detected 17O-MRI is much greater than that of direct 17O-MRI23. The sensitivity increase is far greater than the difference in the intrinsic relative sensitivity of the 1H and the 17O nuclei—a factor of 34. The main factor affecting the sensitivity increase is the fact that, while direct 17O measurement detects only the H217O molecules (0.02 M in natural abundance), the PRIMO method detects protons of all water molecules (110 M). This S/N increase enables the use of the PRIMO method for imaging the H217O distribution in samples in which the H217O concentrations are fairly low. Moreover, the longer T2 relaxation time of water protons enables the use of one-shot ultrafast imaging methods such as EPI, which are not applicable for 17O MRI because of the short T2 of the 17O nuclei. The longer T1 relaxation times of water protons requires long time-to-return between acquired images. However, the relaxation period can be used for multislice image acquisition, and time consumption thus is minimized.

Although the PRIMO signal does not depend on the proton T2, it is proportional to TE. Because the maximum S/N ratio is obtained for TE = T2, it is obvious that the method works better in systems with longer T2. As T2 in biological tissues tends to be shorter at higher magnetic fields, there might be an advantage in performing the experiment at low magnetic fields. The T2 effect, however, might be compensated partially by the increase in S/N ratio at the higher field.

A crucial aspect of the PRIMO method concerning its applicability to human studies is the power deposition caused by the irradiation at the 17O resonance frequency. This issue will be examined in further studies in which optimization of the method will be dictated by an upper limit on the power deposited by the 17O irradiation.

Previous in vivo direct 17O-MRI studies (14) showed that the H217O concentration level in a cat brain after inhalation of 17O-enriched molecular oxygen is ≈2–3× greater than the natural abundance of H217O. This estimate is based on measurements performed on a 1 × 1 cm voxel that represented the cat’s cortex. It is reasonable to expect that local increase of H217O concentration in the area of active oxidative metabolism will be higher than the H217O concentration in such a large voxel. Proton-detected 17O-MRI provides a much higher resolution than direct 17O-MRI and, combined with a fast acquisition method such as EPI, would be able to reflect the increase in local H217O concentration levels. Optimization of the proton-detected 17O-MRI sequence makes it possible to detect such small concentrations of H217O while preserving the advantages of proton MRI.

Acknowledgments

We thank Dr. Seong-Gi Kim for his extensive help with the MRI experiments. This work was supported partially by a grant from the U.S.-Israel Binational Science Foundation to G.N. and K.U. and by a grant from the National Institutes of Health National Centers for Research Resources (Grant RR08079).

ABBREVIATIONS

TE

time-to-echo

TR

time to return

PRIMO

proton imaging of oxygen

EPI

echo-planar imaging

PRESS

point-resolved spectroscopy

DEPRESS

decoupled PRESS

CSF

cerebrospinal fluid

Footnotes

§

Mateescu, G. D. & Fercu, D. 12th Annual Meeting of the Society of Magnetic Resonance in Medicine, Aug. 14–20, 1993, New York, p. 110 (abstr.).

References

  • 1. Thulborn K R, Waterton J C, Radda G K. J Magn Reson. 1981;45:188. [Google Scholar]
  • 2.Ogawa S, Lee T-M, Nayak A S, Glynn P. Magn Reson Med. 1990;14:68–78. doi: 10.1002/mrm.1910140108. [DOI] [PubMed] [Google Scholar]
  • 3.Kwong K K, Belliveau J W, Chesler D A, Goldberg I E, Weiskoff R M, Poncelet B P, Kennedy D N, Hoppel B E, Cohen M S, Turner R, et al. Proc Natl Acad Sci USA. 1992;89:5675–5679. doi: 10.1073/pnas.89.12.5675. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Ogawa S, Tank D W, Menon R, Ellermann J M, Kim S-G, Merkle H, Ugurbil K. Proc Natl Acad Sci USA. 1992;89:5951–5955. doi: 10.1073/pnas.89.13.5951. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Raichle M E, Mintum M A, Hersc P. In: Brain Imaging and Brain Function. Sukoloff L, editor. New York: Raven; 1985. pp. 51–59. [Google Scholar]
  • 6.Fox P T, Raichle M E. Proc Natl Acad Sci USA. 1986;83:1140–1144. doi: 10.1073/pnas.83.4.1140. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Luz Z, Meiboom S. J Am Chem Soc. 1964;86:4768–4769. [Google Scholar]
  • 8.Arai T, Nakao S-I, Mori H, Ishimori K, Morishima I, Miyazawa T, Bernhard F Z. Biochem Biophys Res Commun. 1990;169:153–158. doi: 10.1016/0006-291x(90)91447-z. [DOI] [PubMed] [Google Scholar]
  • 9.Arai T, Mori K, Nakao S-I, Watanabe K, Kito K, Aoki M, Mori H, Morikawa S, Inubushi T K. Biochem Biophys Res Commun. 1991;179:954–961. doi: 10.1016/0006-291x(91)91911-u. [DOI] [PubMed] [Google Scholar]
  • 10.Pekar J, Ligeti L, Ruttner Z, Lyon R C, Sinnwell T M, van Gelderen P, Fiat D, Moonen C T W, McLaughlin A C. Magn Reson Med. 1991;21:313–319. doi: 10.1002/mrm.1910210217. [DOI] [PubMed] [Google Scholar]
  • 11.Fiat D, Ligeti L, Lyon R C, Ruttner Z, Pekar J, Moonen C T W, McLaughlin A C. Magn Reson Med. 1992;24:370–374. doi: 10.1002/mrm.1910240218. [DOI] [PubMed] [Google Scholar]
  • 12.Fiat D, Kang S. Neurol Res. 1992;14:303–311. doi: 10.1080/01616412.1992.11740074. [DOI] [PubMed] [Google Scholar]
  • 13.Fiat D, Kang S. Neurol Res. 1993;15:7–17. doi: 10.1080/01616412.1993.11740100. [DOI] [PubMed] [Google Scholar]
  • 14.Pekar J, Sinnwell T, Ligeti L, Chesnick A S, Frank J A, McLaughlin A C. J Cereb Blood Flow Metab. 1995;15:312–320. doi: 10.1038/jcbfm.1995.36. [DOI] [PubMed] [Google Scholar]
  • 16.Hopkins A L, Barr R G. Magn Reson Med. 1987;4:399–403. doi: 10.1002/mrm.1910040413. [DOI] [PubMed] [Google Scholar]
  • 17.Hopkins A L, Haacke E M, Tkach J, Barr R G, Bratton C B. Magn Reson Med. 1988;7:222–229. doi: 10.1002/mrm.1910070210. [DOI] [PubMed] [Google Scholar]
  • 18.Kwong K K, Hopkins A L, Belliveau J W, Chesler D A, Porkka L M, McKinstry R C, Finelli D A, Hunter G J, Moore J B, Barr R G, et al. Magn Reson Med. 1991;22:154–158. doi: 10.1002/mrm.1910220116. [DOI] [PubMed] [Google Scholar]
  • 19.Hopkins A L, Lust W D, Haacke E M, Wielopolski P, Barr R G, Bratton C B. Magn Reson Med. 1991;22:167–174. doi: 10.1002/mrm.1910220118. [DOI] [PubMed] [Google Scholar]
  • 20.Meiboom S. J Chem Phys. 1961;34:375–388. [Google Scholar]
  • 21.Burnett L J, Zeltmann A H. J Chem Phys. 1974;60:4636–4637. [Google Scholar]
  • 22.Ronen I, Navon G. Magn Reson Med. 1994;32:789–793. doi: 10.1002/mrm.1910320616. [DOI] [PubMed] [Google Scholar]
  • 23.Ronen I, Lee J-H, Merkle H, Ugurbil K, Navon G. NMR Biomed. 1997;10:333–340. doi: 10.1002/(sici)1099-1492(199710)10:7<333::aid-nbm465>3.0.co;2-e. [DOI] [PubMed] [Google Scholar]
  • 24.Reddy R, Stolpen A H, Charagundla S R, Insko E K, Leigh J S. Magn Reson Imag. 1996;14:1073–1078. doi: 10.1016/s0730-725x(96)00227-5. [DOI] [PubMed] [Google Scholar]
  • 25.Stolpen A H, Reddy R, Leigh J S. J Magn Reson. 1997;125:1–7. doi: 10.1006/jmre.1996.1071. [DOI] [PubMed] [Google Scholar]
  • 26.Luz Z, Meiboom S. J Am Chem Soc. 1964;86:4764–4766. [Google Scholar]
  • 27.Luz Z, Meiboom S. J Am Chem Soc. 1964;86:4766–4768. [Google Scholar]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences

RESOURCES